Whether it is positive or negative, I do appreciate it as it helps me to learn and improve the product. I really didn't expect this to get any attention, let alone dozens of comments!
To clarify: This was originally designed to help me auto migrate URL schema. I am learning as I develop this website, and SEO has been one of those vague topics where there are few hard rules. I wanted to leave space for experimentation. As I rolled it out, I became intrigued with how it functions and wanted to share my experiment with you to get feedback.
Based on the feedback, I plan to change the logic such that:
- I will track which URLs are associated with which products
- If user hits 404, I will check if there was previously a product associated with that URL and redirect accordingly
- If it is a new 404, I will display a 404 page which lists products with similar names
I appreciate everyone hopping in to share their perspective!
I am sure everyone is already familiar with Examine. It is perhaps the biggest resource on supplement research. The way they've been operating is that they focus on a few health areas and then scout for research papers that mention related health outcomes, and then describe the link to different supplements. This is a hugely laborious undertaking, but now with AI, many parts of this process can be automated, and that's what I am building.
Here is an example of all research that is linked to Increased Short-Chain Fatty Acid Levels:
I've now indexed thousands of research papers and extracted insights from them that link those studies to different supplements available on the market. The long-term goal is to create the first of the kind supplement store where every supplement on the marketplace is linked to research. Unlike Examine, I intend to make all research summaries public and instead focus on affiliate revenue from the related supplement sales.
The biggest challenge is data accuracy. Due to these being complicated topics, I am avoiding extracting any insights beyond demonstrating a link between the study, health outcome and substance. Users are then linked to the actual research papers where they can building confidence in their decision. However, over time, as AI models advance, I am planning to expand this to a complete insight engine.
I am fascinated by the problem domain, and more generally data normalization problems and applications of LLMs to data normalization problems. I am now doing it as a hobby that I foresee to be a money losing activity for a long time into the future, but I also think there is a fair chance that Pillser could emerge as a preferred website for people who want to buy supplements due to the unique combination of science and inventory.
I am probably a few months away from having the complete database, but wanted to share for early feedback. The website is called https://pillser.com/, and you can already search for different health goals and associated research directly from the landing page.
I am still figuring out the positioning, but the idea is to allow search among thousands of supplements and compare the matching ingredients in a spreadsheet-like interface.
In this example, we see every supplement that contains "magnesium citrate", price of the supplement, totals contents, servings, amount per serving, and price per mg.
This way is the quickest way to figure out which supplement to buy if you are looking for a specific ingredient, as it tells you the price and quantity without needing to browse every product individually.
I've also collected attributes that help to narrow down supplements based on the attributes such as gluten free, vegan, etc. and certifications such as GMP Certified, etc.
I geek out about LLMs and data normalization. I wanted to build something in an intersection of my skills and something that would appear as borderline absurd but have counterparts with billions of revenue (iHerb). I bumped into iHerb and the problem by pure coincidence as I got frustrated trying to buy supplements for myself. It started with a spreadsheet (https://docs.google.com/spreadsheets/d/1qnhEhBPkpI6rY2kdXo5z...) and quickly escalated to what you see today.
so many people helped me a long the way. Especially Remix community. Without their Discord community, it would have taken me twice as long to build this.
Defaults matter. If you are like myself in that 99.9% of the time you want to paste plain text, having that be the default is more useful than having to do a three-key chord.
Why not just use the Paste and Match Style menu item or ⌥⇧⌘V?
- This app can make it the default behavior.
- That menu item is not available in all apps.
- The keyboard shortcut is hard to remember and type.
- In some apps, like Chrome, the keyboard shortcut is different.
- This app can exclude certain apps, preserve links, and also remove tracking parameters from URLs.
Let's be honest, with most people Ctrl+v/Cmd+v is more ingrained in their muscle memory than the longer keystroke combo. More than a few time each week I find myself silently cursing myself for not pasting as plain text.
It is an interesting name, but I would personally not pick it for my project, even if it was in the "dev" space. It feels like branding it would be quite tough.
As described elsewhere, here is how the new logic will work:
- I will track which URLs are associated with which products
- If user hits 404, I will check if there was previously a product associated with that URL and redirect accordingly
- If it is a new 404, I will display a 404 page which lists products with similar names